Need test amp

[tvrgeek]

I would like to have (build) about a 2W, very flat, very wide bandwidth test amp, battery powered if possible. The purpose is for an instrumentation amp for driver measurements out in the middle of a field away from home where the background noise is low enough. Any recommendations? Sounds like a good use for a small chip amp.

[digi01]

tda1521

z

[tvrgeek]

I need a wide power bandwidth, at least 5 to 50K to prevent phase anomalies in MLS measurements, and it needs to be reasonably clean. 100K would be far better. When investigating tweeters, you need to see where the breakup modes and harmonics are.

The spec sheet only shows what it does at 12W, not at half or one where I will be using it. Half a percent distortion is enough to give me issues. For instrumentation, I like to follow the 10X minimum rule. .01 would be much better. I only need a single amp, not twin. I was wondering about some of the publiched op-amp/buffer headphone amps.

[AndrewT]

Originally Posted by tvrgeek

I was wondering about some of the published op-amp/buffer headphone amps.

this seems possible.

[geraldfryjr]

the old ta7205ap is one of my favorites of yesteryear it runs on 12v and is 7watts. .07% thd at 1 watt but not very good above 10khz acordinng to specs,it was used in many car audio boosters in the 80's and was the best sounding chip of the day. also used was the tda2002 (lm383),thatcomes to mind.as i am not up to date on any recent chip numbers.there are lots of chips that are suitable it all depends on your power supply requirements

[lhwidget]

I did exactly what you're thinking about for my SoundEasy set-up. I went with a normal 110 VAC power supply though (I'm measuring inside my apartment).

The LM3886, with only the recommended Zobel on the output got me to 50 kHz @ -0.75dB, and the distortion was very low (looking at loop-back tests in SoundEasy).

An LM3886 on a ±20 VDC switching power supply, powered by a car battery?

I'm not sure what I'm talking about here as I have no experience with switching power supplies (not sure if you can generate ± power rails, or just a higher single-sided DC voltage). Should be good for 10 to 15 watts with very low distortion if the power supply is clean.

[Th3 uN1Qu3]

Switching supplies produce noise, it's best not to have it in the first place than trying to get rid of it afterwards.

I'd use a fast opamp (lots of negative feedback) and slap a class AB stage after it. Question: Do you need it on a single supply or can you use two batteries for a +/- supply? I have a ridiculously simple amp of my own design running on +/-10v (will do well up to 18 or down to 9v per rail, and it outputs 3.5 clean watts per channel).
It's a plain class B stage (exactly TWO transistors per channel) with negative feedback. I use a OPA2134 as preamp and feedback and it does quite well in removing the crossover distortion. Biasing it in class AB should reduce THD to negligible levels.

I can't guarantee for bandwidth however, my hardware test equipment is limited atm, i can simulate it of course, but i can only test the real thing up to 24kHz and down to -90dB noise/THD.

Edit: I measure my drivers at home and my boxes ended up sounding good so far.

[tvrgeek]

I am looking to build a totally battery powered set up. ( one way to defeat the 60 and 120Hz artifacts) I already need 48V for my mic preamp and A2D. That's a lot of 9V batteries, but do-able. I really only need a couple of watts.
I downloaded maybe 20 small head-amps designs but now can't find any of them. Back to google. I think one of the National buffers may have enough power.

Those who use SE know how important the very wide bandwidth is. Otherwise you get bogus phase measurements at high frequencies. 50K is minimum. If I were dreaming, I would pick at least 100K.

[Th3 uN1Qu3]

I understand, you need flat phase response not necessarily bandwidth. Then i must warn you - no chipamp will do that. My recommendation still stays, go with a high quality opamp and a discrete output stage.

[tvrgeek]

Sounds like the plan. I started tracking them down again last night. Several with op-amp inputs and mos-fet outputs. I would love to see your schematic Th3.

The disadvantage in using +/- supplies from batteries is how inconsistent the batteries are and how they drain. Low power circuits are best served with an artificial ground reference. Far lower offset. Safer. Of course, another entire circuit. One could link regulators with a safety shutdown I guess.